Quoting Justin Ruggles (2014-10-30 15:24:45) > On 10/30/2014 03:01 AM, Anton Khirnov wrote: > > It is not supposed to be set by demuxers. > > --- > > libavformat/nutdec.c | 1 - > > 1 file changed, 1 deletion(-) > > > > diff --git a/libavformat/nutdec.c b/libavformat/nutdec.c > > index c3f5f4b..69057e9 100644 > > --- a/libavformat/nutdec.c > > +++ b/libavformat/nutdec.c > > @@ -399,7 +399,6 @@ static int decode_stream_header(NUTContext *nut) > > GET_V(stc->msb_pts_shift, tmp < 16); > > stc->max_pts_distance = ffio_read_varlen(bc); > > GET_V(stc->decode_delay, tmp < 1000); // sanity limit, raise this if > > Moore's law is true > > - st->codec->has_b_frames = stc->decode_delay; > > ffio_read_varlen(bc); // stream flags > > > > GET_V(st->codec->extradata_size, tmp < (1 << 30)); > > > > Anything this would theoretically break? Not saying that line should > stay there, but it would be good to know if something needs to be > handled elsewhere to avoid breaking stuff. >
In general, it could affect the way lavf makes up timestamps. But specifically NUT stores both pts and dts, so this should not be a problem. -- Anton Khirnov _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
